Estimating Wetland Condition Locally: An Intensification Study in the Blackfoot and Swan River Watersheds Prepared for: The U.S. Environmental Protection Agency Prepared by: Melissa Hart, Linda Vance, Karen Newlon, Jennifer Chutz, and Jamul Hahn Montana Natural Heritage Program a cooperative program of the Montana State Library and the University of Montana December 2015 Estimating Wetland Condition Locally: An Intensification Study in the Blackfoot and Swan River Watersheds Prepared for: The U.S. Environmental Protection Agency Agreement Number: CD-96814001-0 Prepared by: Melissa Hart, Linda Vance, Karen Newlon, Jennifer Chutz, and Jamul Hahn ©2015 Montana Natural Heritage Program P.O. Box 201800 ● 1515 East Sixth Avenue ● Helena, MT 59620-1800 ● 406-444-5354 i This document should be cited as follows: Hart, Melissa, Linda Vance, Karen Newlon, Jennifer Chutz, and Jamul Hahn. 2015. Estimating Wetland Condition Locally: An Intensification Study in the Blackfoot and Swan River Watersheds. Report to the U.S. Environmental Protection Agency. Montana Natural Heritage Program, Helena, Montana. 52 pp. plus appendices. ii EXECUTIVE SUMMARY This report summarizes the results of our fourth statewide rotating basin assessment, focusing on wetlands in the Blackfoot and Swan subbasins of western Montana. We assessed wetland condition within nine watersheds at multiple spatial scales. We conducted Level 1 GIS analyses that produced: 1) wetland landscape profiles, which summarize information on wetland abundance, type, and extent within a given watershed; and 2) a landscape characterization, which characterizes the anthropogenic stressors such as roads and land uses, as well as general information regarding wetland landscape context, using readily available digital datasets. We carried out Level 2 assessments to provide rapid, field-based assessments of wetland condition based on four attributes: 1) Landscape Context; 2) Vegetation; 3) Physicochemical; and 4) Hydrology. Finally, Level 3 intensive assessments provided detailed information on the structure and composition of wetland vegetation at a subset of sites. This multi-tiered framework allows for the incorporation of multiple scales of assessment, integrating landscape-level information, ambient wetland condition, and site-specific data. We included all digitally mapped wetlands to produce wetland landscape profiles for the project area. For the Level 1 landscape characterization and Level 2 and Level 3 wetland assessments, the target population included all mapped palustrine wetlands greater than 0.1 ha. We followed a spatially balanced sampling approach to select wetlands for assessment. For Level 1 values and Level 2 assessment scores, we calculated descriptive statistics and assessed the range and distribution of each metric by examining frequency histograms. For Level 3 assessments, we calculated multiple vegetation metrics to conduct a floristic quality assessment (FQA). The FQA accounts for the presence of both native and exotic species, as well as individual plant species’ tolerance of disturbance. We determined the relationships between Level 3 vegetation metric values, Level 2 assessment scores, and stressors recorded at assessment sites by examining Spearman’s correlation coefficients. Based on digital mapping, wetlands and other water bodies within the study area totaled 97,847 acres (39,597 hectares). The majority (76%) of the mapped acres are palustrine wetlands. These totals include deepwater areas such as lakes and river channels, which provide critical aquatic habitat and other valuable ecosystem services, but are not considered wetlands. We conducted a Level 1 landscape characterization of 1,000 mapped palustrine wetland polygons at three spatial scales, examining our recently developed Human Disturbance Index (HDI) within 100-m, 300-m, and 1,000-m envelopes around each polygon. Mean HDI scores were relatively consistent across all three scales for wetland types in the Blackfoot and Swan subbasins, suggesting moderate levels of disturbance. When scores were calculated separately for surveyed and non-surveyed (randomly chosen) wetlands, results suggested a public land effect: 1) randomly chosen wetlands had lower mean HDI scores for public than for private lands, whereas surveyed wetlands – tending to be nearer roads – showed no statistically significant difference between public and private lands; and 2) across all wetland sites, mean iii HDI scores were lower for surveyed sites (which tended to be on public land) than they were for randomly selected sites. We visited 126 sites during the summer of 2013 and 2014, sampling 24 probabilistic sites and 26 targeted sites with the MTNHP Level 3 EIA protocol, and the rest with the Level 2 protocol. The Intermontane Prairie Pothole was the most common system sampled (37 sites). Rocky Mountain Subalpine-Montane Fen was the second most commonly sampled ecological system (26). Other systems encountered in this study included Alpine-Montane Wet Meadow, Western North America Emergent Marsh, Wooded Conifer Swamp, Rocky Mountain Vernal Pool, and several Riparian Woodland and/or Shrubland systems. Level 2 condition scores were calculated for all 126 wetlands sampled. Scores ranged from 52- 100 out of a possible range of 21.5-100. We divided our assessment scores into four categories defined relative to their departure from reference standard. Most sites were either at or near the reference standard, or slightly departed from reference. Thirty-eight sites had no observed stressors in the assessment area (AA), whereas only 17 sites had no observed stressors within the 200-m envelope. No site’s impact rating was categorized as Very High. Impact scores tended to be highest for the Landscape Context and Vegetation attributes within both the AA and the 200- m envelope (Tables 16 and 17). The Physicochemical attribute made the least contribution to higher impact scores, with nearly all sites falling in the No Impact and Low Impact categories. Recreation/human visitation and livestock grazing were the most common stressors potentially impacting Landscape Context within the AA, while unpaved roads were the most observed stressor in the 200-m envelope. Vegetation stressors included browsing by native ungulates, livestock grazing, and beaver activity. Beaver activity was also the most common Hydrologic stressor, along with impoundments. Trash or refuse dumping was the most common Physicochemical stressor. We completed 24 Level 3 intensive assessments within the project area, encountering 309 plant taxa. The average number of species encountered per site was 34 (range 7-76). Of the 282 taxa identified to species, 260 (92%) were native species and 19 (7%) were exotic species. We calculated FQA metrics for all 24 Level 3 assessment sites. Mean C-value across these sites was 4.91 (range 3.44 – 6.34). Most C-values for native species encountered fell between 3 and 8 (Figure 25). Species at the lower end of that range are found in diverse habitats with little to moderate disturbance, while those at the higher end tend to be habitat specialists or have low tolerance for disturbance. To understand the effectiveness of this assessment framework in determining the condition of wetlands in the southeast Montana project area, we compared Level 3 assessment results with Level 2 assessment results. Impact ratings within the AA and within the 200-m envelope around the AA showed moderate correlations with overall Level 2 assessment scores. The Landscape Context and Hydrologic attribute scores were most strongly correlated with overall impact rating for both the AA and the 200-m envelope (with r values ranging from 0.35-0.54). There were 16 vegetation metrics evaluated in the FQA. Most of the vegetation metrics showed some degree of correlation with either stressors or overall wetland condition, but none of the iv correlations was strong. The strongest correlation was observed between mean C-value of native species and overall condition score. Non-native species richness was negatively correlated with stressor impact scores and overall condition scores, meaning that as impact and condition scores increased toward their maximum values (indicating reference conditions on the ground), the number of non-native species decreased. FQA metrics that were correlated with overall condition scores also were correlated with one or more individual Level 2 attribute scores (Table 24). Not surprisingly, nearly all FQA metrics showed moderate correlation with the Vegetation attribute. Again, non-native species richness was negatively correlated with all four attributes, meaning that as EIA attribute scores increased toward 100 (reference conditions), non-native species richness decreased. Results from this project indicate the wetlands in the Blackfoot-Swan area of Montana are in good to excellent condition. Except for unpaved roads, which are common throughout the study area as a legacy of logging, wetlands are relatively unimpacted by human stressors. However, these unpaved roads often act as sites for weedy species to establish, threatening wetlands and their adjacent buffers. The rotating basin assessment was complemented by an analysis of data from this and previous projects to evaluate whether water permanence is a predictor of wetland condition. Analysis of the National Wetland Plant List for the Western Mountains and Valleys against the Coefficients of Conservatism (C-values) for Montana wetland plants shows that the mean C-value for OBL and FACW species is 5.83,